DESCRIPTION: Recent clinical studies have demonstrated that hypoxia is a strong predictor of local recurrence in cervical carcinomas and metastasis in soft tissue sarcomas. Other studies have associated hypoxia with poor response of head and neck cancer to radiation therapy. One of the most clinically relevant technologies to identify the presence of hypoxia in individual tumors is the reductive binding of 2-nitroimidazoles to cellular macromolecules. The principal investigator and co-principal investigator's laboratories have developed the 2-nitroimidazole drug EF5 as a hypoxia 'marking' agent. Working with collaborators, they have shown that monoclonal antibody-based quantitation of EF5 can predict radiation resistance in multicellular tumor spheroids and individual rodent tumors. The studies proposed herein will determine whether EF5, when given to humans, is safe, efficacious (i.e., binds to hypoxic tumor cells) and can be used to predict treatment outcome and tumor biology (metastasis and recurrence). They will perform this clinical trial in patients with high grade truncal and extremity soft tissue sarcomas and cervical carcinomas. The optimal intravenous EF5 dose will be determined by considering minimal toxicity to the patient and optimal signal contrast of the antibody-based measurement of EF5 binding (specific aims 1.1 and 1.2). The signal contrast will be determined in vitro based on EF5 binding kinetics of cells and tissue from individual patient tumors. In specific aim 1.3, they will compare the in vivo tumor oxygen distributions based upon EF5 binding with those determined with Eppendorf needle electrodes. In specific aim 2.1, they will determine if power Doppler ultrasound can be used to guide biopsies to regions of low blood flow and therefore, hypoxia. They will also determine the threshold number of biopsies, in order to accurately support those of specific aim 3; e.g., to determine the predictive strength (for relapse) of EF5 binding, in combination with tumor and patient characteristics, Eppendorf electrode measurements and tumor thiol concentrations. In both high grade truncal and extremity soft tissue sarcomas and cervical carcinomas, separate multivariate analyses will be performed.